Federated Annotation

Objective

Federated learning requires that the different datasets share the same labels in order to train models separately and then aggregate.

Add dataset

Click on the Add Dataset button in your project to add a new dataset. Load the data from the new institution or entity.

Gather multiple datasets from different institutions into one project

Annotate on multiple datasets with the same set of labels and same users

Export annotations for use in a Federated Learning model

Each dataset is trained separately with its corresponding annotations and the models are aggregated for validation or additional training cycles.